 Live from the Moscone Convention Center in San Francisco, California, it's The Cube at Oracle OpenWorld 2014. Brought to you by headline sponsor Q-Logic with support from HGST, Violin Memory, and MarkLogic. Now, here is your host, Dave Vellante. Hi everybody, we're back at Oracle OpenWorld 2014. We're here at Moscone South. Inside the Q-Logic booth, our friends at Q-Logic have sponsored us now for the fifth year in a row at Oracle OpenWorld. It's going to tunnel our way in in 2010 and we like it here. So we just sort of embedded ourselves like a tick. The Cube does that. So The Cube is a SiliconANGLE's live mobile studio. We go out to the events, we extract the signal from the noise. Steve Karam is here. He's the Oracle Alchemist, Cube alum. He's now with Delfix. Steve came on last year and we were rippin' about what it's like to be an Oracle DBA. He's the director of customer education and experience at Delfix. Really interesting company. One that I didn't know much about until Steve and I started talkin' off camera. Steve, welcome back to The Cube. It's great to see you again. Thank you very much, Dave. Good to see you too. So let's get right into it. Let's start with Delfix because it's an interesting story. If I can, just as a point of reference, a lot of the people in our community know about Actifio. Actifio does copy data services. A bunch of folks have been following them. A hot company in the East Coast. You guys are similar but different. So I'm sorry to start with sort of a reference point of another company but it helps sometimes with our audience and in my simple brain to do that. Talk about Delfix, what you guys do and how you're different from the other copy data service guys. Sure. So copy data is at the core of what Delfix does. If you look at what Delfix does, we take source databases, source data and we bring it and sink it into Delfix from there we can provision out virtualized copies of apps, databases, whatever we want in minutes. Well we do sync to over 10 different heterogeneous sources, flat files, Oracle databases, SQL server, Postgres, et cetera. The big difference is it's not just copy data. Copy data is great. It has to be at the foundation of that. But it's the layer over top to solve the business challenges. It's the ability to provision out multiple copies over time, snapshot data over time to be able to provision your entire app stack as of a certain point in time like the end of a fiscal year, data masking, various other capabilities that really round out that entire cloning concept. Okay, so it's based on snapshot technology. We talked again off camera, you've got a catalog which is fundamental to anything like this. So the catalog is a metadata catalog that tells me what's been copied, where it is, what version of it is. So talk about the architecture a little bit more in the secret sauce if you would. Sure, so at the architectural level, you basically got Delphix which is a software appliance. It can live in a vSphere environment. It can live in Amazon EC2. It can live in VMware vCloud. And it can run on any storage. It can run on any hardware, any Intel based hardware as far as that's concerned. And it syncs to the source systems. Those source systems can pump in change data over time. It builds what we call a time flow. So you're asking about a catalog. What our catalog looks like is almost like the cover flow in iTunes. You've got a catalog of all of your sources with all of their changes over time. You can bookmark and snapshot those points and then you pick a point in time, you click the provision button and you have a full copy of that environment in a few minutes with no extra storage. And that's where the change data part comes in. But that environment that you've cloned can then also be, it has its own time flow. It can be rewound. You can provision virtual databases off of virtual databases. You can build a golden copy that you then replicate to another Delphix engine. And you provision masked copies off of that. And that's just the database. That doesn't include the app stack, the other databases that support it, flat files, all of that. And that's where Delphix brings its value. Is it encompasses the enterprise? It encompasses all of the needs you might have. So one obvious case is data protection. Is that right? We'll have to talk about it. So the company was founded by Jetta Daihua who was the founder of Avamar, sold Avamar to EMC. So he knows a little bit about backup. So that's an obvious use case, but it transcends backup. And I want to get to those other use cases. But what a lot of customers do today, they'll go out and they'll buy a purpose-built backup appliance. And they'll, which is great, de-dupe, wonderful, relative to tape. But then what happens when you run out of capacity is you buy another one. And you buy another one. And when you buy another one, in your architecture, you're using very space-efficient snapshots and you're just saving the changes. You're using your catalog. So you need a lot less capacity, less capacity, less hardware, less software, less maintenance, less people, less op-ex, all that wonderful TCO stuff. Is that where you started or no? That is where we started. I wouldn't say it was more from an application agility framework. And we've expanded from there into data protection as an obvious use case, as you mentioned, and then other use cases. Again, coming from that point that we can store it in a very small footprint, we can grow that vertically or horizontally. And you can build, again, it was an agile application development use case at first and we've expanded drastically into other markets. That's interesting. So you started in application development because the problem with backup is it's insurance. Yes. Nobody wants to pay a ton for insurance. And so if you keep pouring money into the insurance policy and not getting anything in return, people get mad. Yep, absolutely. Okay, but so then the architecture that you're talking about lends itself to other use cases. So let's talk about those. Application development, test dev is an obvious one. Absolutely. Populating data warehouses and analytics is maybe another one. So I wonder if we could just sort of progress through those and talk about those a little bit. Sure. So as far as analytics is concerned, we've got some amazing use cases that are going on right now with some of our customers where they're actually provisioning out a virtual database, a copy of many terabytes. You can provision out a clone of your entire Oracle data warehouse. Doesn't matter how big it is. And give one to each reporting team. You can give copies of production with no overhead on the source system to each reporting team. You can build complex analytics because it's a virtual database. It's fully read right. So what they can do is take their source production, source data warehouse. 10 minutes later, you've got a clone of each. Do whatever destructive changes and testing you want on it at the end of your testing, at the end of your analytics, whatever you're trying to work out, you can rewind, you can trash it, you can keep on using it. So there's a lot of cool features in there that lend to building highly complex analytics scenarios. So you think of like an actuarial company where what they're going to want to do is take their source data, clone it off for analytics purposes and run risk models, run assessments, rewind, run them again and try different formulas to expand their own business. Mega what if scenarios. So let's unpack this a little bit because what you're saying sounds good but a lot of people might not fully grasp the nuance. So let's take the world pre-delfix. Let's take a test and dev example. I want to spin up an instance, do some test dev. Take me through what I have to do, pre-delfix with the data sets and some of the challenges that I face. Sure, you know I can do that because I lived it for 15 years, 16 years. I'll give you an example at a place I recently worked. Database was eight terabytes in size. They wanted five QA and dev environments, a total of five. It took three days to clone our source database over there. So you think about that. Those five environments, it's going to take three days per environment to clone it or refresh it. It's going to take 40 total terabytes to populate them of storage and it's going to cut into every timeline you've got. A single error just kills the entire development lifecycle. So what we had were teams that didn't want to give up their QA box for three days to have this happen and possibly screw up. You add, but we needed constant refreshes. So I'll tell you what we ended up with. We ended up with a two year old development system that hadn't been refreshed in two years. We ended up with data subsets which was what we brought in about 500 gigs worth and hope that was enough, you know? Because the amount of time it would take for the DBA, CIS admins, the storage requirements, the project managers, all the way up to the managerial level, every time you needed to refresh, we're talking about tens of thousands of dollars in waste and man hours and storage and time and downtime for your environments. And then when that didn't happen, if a company, if we decided, well, let's go ahead and not refresh. Let's just use what we've got. Now you've got errors that make their way into production. And as you know, downstream errors cost a heck of a lot more. Once they get there. So you were essentially doing sampling with old data. Yeah, exactly. For your QA. For QA development, whatever else. But that is so common. Which is common. Everybody does. Everybody does it. Now bring in Delfix. How is it different? How is your life, my life as a developer, different with your platform? Sure. So with Delfix, you basically, you've got the Delfix appliance, which it's a virtual data platform. As I mentioned, it runs in vSphere. You install Delfix and you sync up your data sources with it. Not only that, but we have a product called Jetstream, which is a developer and end user oriented provisioning tool. It has branching. It has bookmarking. It has rewind. It has rollback. You can almost think of it as like Git for all of your data sources. Awesome visual interface, developers, QA engineers, they go in, they look at the branch of the software development cycle. They pick a point in time and they say, give me that. If they want to do some destructive testing, they say, save this branch, rewind, or switch over to this one. And Delfix will automate and orchestrate. That's the biggest part. The bringing down, bringing up VDBs, DBAs don't have to lift a finger. Once they're set up for self-service, Delfix will automate the bringing down, the bringing up of the Oracle VDB. It'll mount file systems and do all the heavy lifting. End users just click the button and they're off. And again, another nuance here that I want people to understand, Steve, so help us is the freshness of the data. Why is operating on fresh data so important? Maybe it's obvious to people, but it wasn't to me when I first started thinking about this. Sure. Well, it's absolutely important from a performance aspect. Let's talk one aspect of QA. There's functional QA and there's performance QA. From a performance level, you need to have your full range of data because all it takes is an extra 1,000 rows in production or an extra 10 megs. And suddenly, yeah, your performance, it's on a nice clip and then it just drops. So being able to test with a full set is very important. From a functional aspect, you need to have the most recent cases for whatever your business does. Your business just started adopting a new use case. They sell a new product. They have a new line in their inventory. You need to have that fresh data for QA purposes or you're not going to find the errors at the right time. Now, a lot of clients that I've worked with in the past, their legal department, the GC, doesn't want them keeping data around forever. Can you help me with that problem from a compliance standpoint? Can you help me defensively delete data? That's actually a really interesting one because you've got some that don't want to keep data around and you've got plenty that want to keep it around for a long time. And they want to categorize it and keep track of it. We can help with both. So basically with Delfix, we can keep data for as long as you want. We set our own retention windows. Delfix is implicitly a right once read many type of environment. So we can snapshot data, for instance, at the end of each fiscal year over the course of three years, a company can then provision a clone of their entire app stack as it was at the end of fiscal year three years ago or at a specific point in time for legal, for compliance reasons. A company's ever called a court, you know? We can help with that. But as far as not keeping the data around, fully adjustable. So the timelines, the retention, what we keep is fully adjustable. We can get rid of virtual databases as of certain points in time. And then we also have masking capabilities so that the consumers of those virtual databases don't get the actual data. They get a garbled protected version of the actual data. Still fresh, still on time, still the same size and smell and feel, but the data, the sensitive data is going to be masked. It's safely masked. Yeah, because I mean, you're right. Some people want to keep it forever, especially these days and want to do analytics on it. And then you got the example of a pharmaceutical company who has a work in process. Exactly. And they were maybe testing, hey, there could be a problem with this drug. Yep. And then later on they figured, oh no, there's not a problem with the drug. And but they have that work in process that's hanging around. I want to delete, if I'm the GC, I want to get rid of that. I want to have a policy to get rid of that. You're saying you can help me with both of those. Yes. Okay. Wow, anything else you want to tell us about Delfix? Because I want to start, I want to switch hats and talk about Oracle, but any other sort of final thoughts on Delfix? Well, you know, so the only thing I would say is like I mentioned, we do have a multitude of products when it comes to that. We are a virtual data platform. You know, we go from modernization, which includes automating the transition from Solaris to Linux. We have a compliance engine. We have, you know, and then we also have a new copy data engine purely meant for the core use case of, I have a bunch of data. I want to sync it. I want to copy it. I want to be able to clone it. So, multiple things and... Cool. Yep. And you guys based here in Menlo Park. Absolutely. To say Jed is your CEO and founder. And we'll be watching. I got to learn more about this company. Okay, let's switch gears. Oracle in transition. Did you hear Larry's keynote last night? I heard of Larry's keynote. I was in flight at the time. Okay, so a lot of talk about cloud. We heard last year was the announcement of 12C. Now we're hearing it's close to actually becoming, you know, widely adopted in C. So, but cloud, cloud, cloud, cloud, cloud was really the message last night. Interestingly, Larry said that we made a commitment to our customers 30 years ago that dictated that we had to get into all three layers of cloud, infrastructure as a service, platform as a service, and software as a service. He never really explained what he meant by that commitment, but nonetheless, they're doing it. And they're all in. So, the company's in transition, its license revenue is down, the on-premise license revenue, its cloud business is exploding. I think it's a two billion dollar run rate, I want to say. It might even be bigger than that. I got my, I'm lost in my notes. Yeah, it's a two billion dollar run rate in all cloud, so pretty substantial, growing at 30% a year. Mark Hurd said they're going to grow even faster, so they're on an accelerated growth curve. He said 40% going forward. What do you make of Oracle? People say they were late to the cloud game and yet their cloud business is exploding. People say they're living off of maintenance, which they are, they're on-premise license businesses declining, their maintenance business is growing. People like Workday say, hey, we're cleaning up against Oracle, Oracle says we have more customers, we signed more new customers this quarter than Workday has since their inception. What's going on? Right, so I do think they were a bit late to the cloud. Oracle's usual response, and this is maybe a little cynical of me, but let's add a letter to the end of our version number and hey look, we're cloud ready. But as we know, that doesn't mean they're cloud ready any more than the eye on Oracle 9i meant that they were internet ready because Java was used. That being said, there's some great innovations coming and they do need it. Like you said, the on-prem dollar that they're earning from that is all tied up in maintenance. I think I mentioned that last year when we were talking about this, is that the maintenance costs are the core of that. Cloud is the way they've got to go. From what I understand, there's going to be some great innovations. I've heard talk about different ways to get to the cloud and do things in the cloud. I still don't feel like I've heard the meat that makes me feel like it's going to be an amazing innovation over other companies right now, but I'm looking forward to seeing it. Well, it's certainly broad. The portfolio is as broad as any software company. We can agree on that, I think, right? That's fair. Absolutely. The fact that they're doing infrastructure and database and middleware, including Java, that's a goodness, because they've got the number one database in the world and Java's the number one programming language in the world. And they've got eight zillion apps that are sort of, to me, still stovepipe, but Fusion apps sort of brings those together. Right, right, so if you're RedStack, the value proposition is I want to test with you is, like you said this last night, Larry, with one click, I could take my on-premise database, make it cloud-ready, and modernize it. He had a great statement. We can bring your apps to modernity, endow the apps with modernity. It's what he said, I was like, wow, what does that mean? But I think what he meant is it's in memory, it's probably compressed, and it's cloud, it's multi-tenant. Sure. One click, I'm skeptical. Well, you know, I can't talk too bad about it because Delphic's promise is, with one click, we can provision your things to the cloud too. But as far as that's concerned, you're right, he's got all the stack, he's got the entire stack from top to bottom, which is great. The hardware side, he also mentioned something about how hardware and software engineered together is the key, and in some ways I agree, I use a Mac, you use a Mac, I notice. And there's a reason we like that, it's because it's all done in lockstep and it all comes out very smooth. So I think that's going to be a good offering, it's going to be a strong one. I'm still, again, maybe it's the skeptic in me, I'm still a little skeptical, I want to see how fast that ramp up really is. I don't know about that one click thing either. I like to think it's going to be that simple, but it's not as easy as it sounds. Any more than ungodly speeds of last open world are as simple as just turning on a new feature and suddenly all of your BI works at ungodly speeds. I think it's back to nine I. What can we learn from that in terms of Oracle's track record on accommodating new innovations? Larry said, I want you to help us stink test this. He said, back in the mainframe days, he said Oracle version two, which was really Oracle's first database. He didn't name it version one because he said nobody would buy version one. No one would buy version one. He was running on a micro, I'm not a micro, it's a Vax, I guess. Deck mini computer, I'm not sure if it was a Vax. Probably was, but anyway, it ran on IBM mainframes. And when the world went to client server, he said, customers came to us and said, we want to port our applications to client server. And Oracle said, well, great, rewrite them. And they were like, we don't want to rewrite them. So we did the hard work to move everything to client server. Now, PeopleSoft did a lot of that. They were one of the first, actually, you know, Oracle owns PeopleSoft. So when he says we, he could be including that in that statement. And he said, now with cloud, we're doing the same thing. So that was sort of Larry's version of history. How accurate is that? How, what can we learn from that in terms of Oracle's ability to seamless, seamlessly? It's not seamless, we all know that. But to seamlessly enough, enough being before you lose the customer, migrate people to the cloud. What can we learn from previous experiences? Well, so from a technical point of view, the reason for skepticism on this is you look at 9i, Oracle 9i, actually it was in Oracle 8 that they added the i to the end, meaning internet. And what did that mean? It meant that it could plug into Java. You could get some advanced functionality with internet by use of Java, Java stored procedures, things like that. 10g came along and they attached a g. The g stood for grid, which no one I don't think really understood at the time. It was the idea of almost basically making your own private cloud inside of your data center. So, you know, I have a running joke on my blog where I scratch out the i and the g and put a c and I say we meant cloud the whole time, you know? Internet, I mean grid, I mean cloud, you know? We've been building these features, but one of the big new features I've heard about in 12c, well, we now have JSON support built into Oracle. Now, in 2014, how long has the apps world been focused on JavaScript development? I mean, we've got all of these technologies that are based on JS, everything. You know, we've got our Mongos and our meteors and our expresses. JavaScript is the foundation for so many fast market apps and yet we're finally supporting JavaScript natively inside of Oracle. So, that's the brilliance of Oracle's marketing. They'll relate to the game and then go and opt into the act like they invented it. They do a really good job. They do a great job at that, though. So, okay, so what do you think happens to Oracle? Obviously, you know, it was interesting. I mean, a lot of people have said, oh, big data's going to disrupt Oracle. Larry's talking about big data. Cloud's going to disrupt Oracle. Larry's up there talking about cloud. Amazon, you know, Microsoft, Google, he said we're going to price the same as Amazon, Microsoft, and Google. We're just going to be more reliable and more secure. I'm not sure if he's including database licenses in that pricing statement. But I heard there's changes coming. Say it again. I've heard there's changes coming in the way we look at Oracle licensing, so that's interesting. Well, it is interesting. And maybe they've looked at it and said, okay, run some thinking or models and say, hey, we can actually get more business. We can make it up in volume, so to speak. And Oracle potentially has the capability to do that. So what do you make of those statements? What's the future of Oracle? So one of the things we talked about last year, I think there's always going to be the enterprise user, the enterprise corporation, all of this cloud. There's companies that will never move to the cloud. Or if they do, it's going to be their own version of it. You know, I think there will always be a home for the enterprise app, for the enterprise system, the enterprise server. The thing that Oracle is doing right is they've finally gotten away from the, we're Oracle, you do what we say, mentality. Big data is getting to be not a problem, but a big thing. Let's go ahead and interface and work with Cloudera. Let's bring other companies into the fold. Let's start opening up our technology and working with others. I think that's going to keep Oracle relevant. If they can change their licensing model to make them more accessible and make it easier to get into in the cloud, I think they could be very, very successful in a way that they're not used to now. And I think that Larry's recent transition to where he can, his Iron Man moment, where he basically handed the company over to Pepper Pot so he could focus on being Iron Man, right? He was in that film too, Larry Ellison. This was his moment where he can now innovate without worrying about being at the brunt of everything and taking care of the finer details. And I'm hoping, I'm hoping that that innovation is along the lines of let's change our business model to support massive innovation and growth from the enterprise as a whole, not just from those big whale companies. So that's interesting. I mean, it's sort of the competitive landscape changes. So what Safra talks about going from a license model to a maintenance to one where Oracle's providing the hardware, the middleware, the applications all in one. So they're doing more for the customer. So Oracle's making more money. The value proposition for the customer has to be we can do this more efficiently. It changes the competitive dynamics. You know, of course, Oracle buying Sun changed a lot of competitive dynamics too, but who's the competition? It's clearly Microsoft, IBM, to a great extent Amazon now. Oracle doesn't want you running its database even though it will let you in Amazon. It wants you running on the Oracle's cloud. So the competitive landscape's changing. And then of course SAP, you know, Larry was hammering Hannah last night. Hannah powers the cloud. What cloud? Who's cloud? Where? You know, who's classic Larry moment? So you got the traditional competitors, many of whom rely on Oracle databases like Salesforce. He pointed out everybody were Orc Day and IBM by the way, IBM, I think buys a lot of Oracle, although maybe some, maybe not internally, but their customers do. But who are the competitors of Oracle now? Well, so I would say that Oracle is doing the equivalent of, you know, fighting a land war in Asia sort of thing. They're enemies of everybody. Oh man, because what you've got, you've got all of these database vendors that they've always been competitors with. Then you've got these little upstarts, the no sequels and the other, you know, the free ones, the open source ones, which have their own use cases and they're disrupting things left and right for big RDBMS vendors like Oracle. They've got, not enemies, but you've got competitors in the apps area and the reporting area, now in the hardware. And then as far as the cloud is concerned, like you said, Amazon, you've got your rack space, you've got your various different clouds. And I do find it very funny that the whole idea of cloud is let's get rid of this lock in and let's get into the cloud and everything. Well, now we're going to be locked into a particular cloud, right? So it doesn't stop there. Oracle is going to try to bring customers into their cloud. They're going to be competing with Amazon big time. I think that's their biggest one. I know Google's expanding their cloud offerings. I don't think they'll compete much, but Amazon has that relational database service, which is tailor made to run Oracle in Amazon's cloud. And from what I understand, it's fairly popular option. RDS is a popular option. And of course Amazon's saying, okay, BYO software, but ideally we want you to put the software in the marketplace, which I would think Oracle doesn't want you to put the software in the marketplace. So very interesting dynamics going on. All right, Steve, I'll leave you with the last word. You can pick Delphix, you can pick Oracle Open World, you can pick your blog and the Oracle Alchemist, but the bumper sticker on the back of the truck pulling away from Oracle Open World this year, 2014. Bags are all packed. The conference center in Moscone's shut down. They've kicked us out, which they'll do on Wednesday. What's the bumper sticker in the back of the truck say Oracle World 2014? Oh man, it's going to probably say Oracle World 2014 time to go to the cloud. I mean, that's what it all is. It's all about being cloud ready and getting there. I will add to that. You said I can, you know, this is the final word. I will say one of the things I mentioned about Delphix specifically, one of the core features I see is that ability to get from your data center to the cloud. That whole click of a button thing that Oracle mentioned sounds great, it does. But how are you going to test it first? How are you going to get there and have a full environment in the cloud, ready for testing, ready to test your entire app stack against to make sure you're ready for that move. Because it's not the kind of thing you just want to jump right into and trust that it's going to be great. Good advice from the Oracle Alchemist. Steve Karamp, thanks very much for coming on theCUBE. It's really always a pleasure. Thank you, Dave, it was great. All right, keep it right there. We'll be back with our next guest right after this. This is Oracle Open World 2014 and this is theCUBE. We'll be right back.